Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
    zoq
    @marcusedel:matrix.org
    [m]
    Did you already order a new hdd or are you now going for a new device?
    1 reply
    Gopi M Tatiraju
    @heisenbuug
    In file included from /home/heisenbug/heisenbug/ncf/mlpack/src/mlpack/methods/cf/ncf_main.cpp:21:
    /home/heisenbug/heisenbug/ncf/mlpack/src/mlpack/core/util/mlpack_main.hpp:53:4: error: #error "Unknown binding type!  Be sure BINDING_TYPE is defined if you are " "including <mlpack/core/util/mlpack_main.hpp>.";
       53 |   #error "Unknown binding type!  Be sure BINDING_TYPE is defined if you are " \
    Where do we have to define BINDING_TYPE?
    I dont see it in any other examples...
    shrit
    @shrit:matrix.org
    [m]
    Not hard, usually HDD are easily accessible in laptops
    rcurtin
    @ryan:ratml.org
    [m]
    1 reply
    anyone have any thoughts on the "one-pager" that NumFOCUS prepared for us?
    I've attached it here
    NumFOCUS is asking for feedback if we have any---to me it looks pretty good at first glance. I thought I would send it to everyone here so that we can get more feedback if folks have opinions :)
    (for context, this is something that NumFOCUS posts on their website and hands out to companies. whether a company actually jumps at the "project needs" and offers funding... I don't think we can count on anything there, but it's nice to have the advertising materials)
    jonpsy
    @jonpsy:matrix.org
    [m]
    wait... NASA uses mlpack?
    jonpsy
    @jonpsy:matrix.org
    [m]
    looks good to me. A nitpick from my side would be to order the "Project Needs" in order of urgency. So, C++ lead engineer and GPU support should be at top, whereas technical writer should be somewhere below.
    shrit
    @shrit:matrix.org
    [m]
    The second page, they added both the generic and detailed needs, I would remove both of the first two lines and keep the last four. Otherwise, it would show that the project requires a huge amount of support
    SuvarshaChennareddy
    @SuvarshaChennareddy
    hello everyone
    im trying to implement the ROC AUC metric for mlpack and Im almost done.
    but I have come across a problem
    the other metric classes have first classified with the data and then computed the corresponding metric
    with ROC AUC I don't think its a good idea to classify within the Evaluate method
    because the models to be tested have different methods for classification
    what should I do?
    SuvarshaChennareddy
    @SuvarshaChennareddy
    Should I leave the classification out of the evaluate method?
    Gopi M Tatiraju
    @heisenbuug

    mlpack/mlpack#1454

    Why is memory check failing? Am I doing something wrong, or will restarting the tests will work?

    2 replies
    shrit
    @shrit:matrix.org
    [m]
    no this is an issue with Jenkins itself.
    Gopi M Tatiraju
    @heisenbuug
    Oh okay, then I think I am done with this pr, coz I was able to build mlpack successfully locally, but I didn't ran the tests
    So what's the solution here?
    And whats causing the issue?
    shrit
    @shrit:matrix.org
    [m]
    I have no idea, maybe zoq do you have an idea what is happening with the memory check issues?
    Thank you very much I am going to give it a review of the builds are passing
    Gopi M Tatiraju
    @heisenbuug
    Also, I am not sure but should I add my name to the list of authors? It's not like I actually made changes to the core logic...
    shrit
    @shrit:matrix.org
    [m]
    Up to you, I would not add my name unless If I add features or modified extensively the file.
    Gopi M Tatiraju
    @heisenbuug
    Yes, I don't think I should.
    Thank you.
    I will get on that issue which you opened, I think I might be able to complete it tonight.
    I would love to see mlpack becoming header only
    Gopi M Tatiraju
    @heisenbuug
    Regarding cosine_tree, there is a part where we are iterating boost::heap::priority_queue but we cant do the same for std::priority_queue
    I mean we can't iterate over them using an iterator, we have to pop the elements
    And since these functions are not modifying the queue anyways should I pass them as value and then pop them, so that it won't mess with original?
    jonpsy
    @jonpsy:matrix.org
    [m]
    say4n: would you be free for a meet today?
    jonpsy
    @jonpsy:matrix.org
    [m]
    zoq: hey, any progress on the thing I mailed you about?
    1 reply
    jonpsy
    @jonpsy:matrix.org
    [m]
    Can someone explain what E_sub(s,a,w) mean?
    say4n
    @say4n:matrix.org
    [m]
    jonpsy: E_{s,a,\omega} is basically the expectation over all states, actions and preference vectors. You can expand it to three integral over the three respective continuous variables.
    jonpsy: yes, I can hop into the meeting and listen in however I’m afraid I couldn’t go over the PR. :(
    say4n
    @say4n:matrix.org
    [m]
    jonpsy: you there? Marcus & I are at the zoom room :)
    1 reply
    jonpsy
    @jonpsy:matrix.org
    [m]
    I forgot to tell you about my end sem. But the updates are: The replay method is done as well
    jonpsy
    @jonpsy:matrix.org
    [m]
    So I was planning on verifiying each functions are running properly & then we can go for checking the main loop. This will be exhausting, but neccessary. Then we have to work on setting up test framework.
    For mario, we have to implement MoA3C (a multi-objective variant of A3C), we'll skip that atm. The other tests are fairly simple.
    So the PR is a pass from eagle's view. I will mark it as ok for review
    Requesting PR reviews, thanks
    say4n
    @say4n:matrix.org
    [m]
    this is quite an amusing short read: https://belaycpp.com/2021/11/24/is-my-cat-turing-complete/ :)
    1 reply